Your AI is a spotter. You still have to lift. Did I confuse you with that one? G
By Cjhaugen · October 29, 2026 · Curated by George's Blog
Your AI is a spotter. You still have to lift.
Did I confuse you with that one? Good! Have a seat and allow me to clarify. The tools we are using today are getting really good. At the rate it’s going, it feels like every day there is something new or improved with what we are using. With Claude connected to our data through MCP, work that used to take an afternoon now takes minutes. And the output keeps getting better when you use it right.
But I saw a phrase recently from Destaney Wishon in a recent post that stuck with me: AI is also “amplifying ignorance”. I couldn’t agree more.
If you don’t know what good looks like, AI will hand you something that looks good. Fast, confident, sometimes wrong, and if you stop checking, you stop noticing.
Not long ago, I ran a budget pacing analysis with the help of an AI input. The output said bid reductions were why we’d started falling behind. At a glance, it looked right. But I knew bids hadn’t changed when the shift happened. So I stepped in. The real cause: a price change that drove CTR way down and likely cost us ad rank.
Thinking is a muscle. Like any muscle, it weakens when you stop using it. You may not notice the ignorance right away, but when you go to work that muscle more independently after not using it to the same degree like you once had, it will hit you like a ton of bricks.
Here’s how I’m trying to keep mine in shape, and or better yet, growing it:
🔥 Warm up before you prompt
Take a few of the minutes that you’re saving to write down what you expect the answer, or other output to be. Then compare the two. The gap is where the learning happens, and it loosens you up to tackle your full set.
👀 Spot-check your reps
Pick one number, claim, and source, and verify it at the source. Every time. If a ROAS figure looks too good, pull the raw report. Was the lift too easy? Or did it feel just right after those last set of reps?
💬 Ask “why,” not just “what”
Make the AI show its work: the query, the logic, the assumptions. If you can’t explain the output to a client without the tool open, you don’t own it yet. You don’t own your muscle growth if you aren’t picking up the weight.
🏋 Do some sets unassisted
Every so often, do a task from start to finish on your own. Build the analysis. Write the first draft. It keeps your baseline strong, so you can tell when the AI output is off. You don’t always need a spotter, and you owe it to yourself to prove that the assisted lifts your taking are helping you grow, not plateau and rely on others
AI is an amazing spotter. But a spotter doesn’t lift the weight for you. The goal isn’t to use AI less. It’s to walk out of every session a little stronger than you walked in.
How are you keeping your thinking muscles in shape?👇